Microarray gene expression classification with few genes: Criteria to combine attribute selection and classification methods

Microarray data classification is a task involving high dimensionality and small samples sizes. A common criterion to decide on the number of selected genes is maximizing the accuracy, which risks overfitting and usually selects more genes than actually needed. We propose, relaxing the maximum accuracy criterion, to select the combination of attribute selection and classification algorithm that using less attributes has an accuracy not statistically significantly worst that the best. Also we give some advice to choose a suitable combination of attribute selection and classifying algorithms for a good accuracy when using a low number of gene expressions. We used some well known attribute selection methods (FCBF, ReliefF and SVM-RFE, plus a Random selection, used as a base line technique) and classifying techniques (Naive Bayes, 3 Nearest Neighbor and SVM with linear kernel) applied to 30 data sets involving different cancer types.